1

The application of infrared imaging and optical coherence 1

tomography of the lacrimal punctum in patients 2

undergoing punctoplasty for epiphora 3

4

Hannah M. Timlin, BSc, FRCOphth,1 5

Pearse A. Keane, MD, FRCOphth,2 6

Geoffrey E. Rose, DSc, FRCS, 1,2 7

Daniel G. Ezra, MA, MD, FRCOphth1,2 8

9

1Lacrimal Clinic, Moorfields Eye Hospital, London, United Kingdom 10

2NIHR Biomedical Research Centre for Ophthalmology, Moorfields Eye Hospital 11

NHS Foundation Trust and UCL Institute of Ophthalmology, UK 12

13

Financial support: None 14

15

Correspondence and address for reprints: 16

Mr Daniel G. Ezra, MA, MD, FRCOphth, 17

Adnexal Office, Moorfields Eye Hospital London EC1V 2PD, United Kingdom. 18

Tel: +44(0)207 253 3411 19

Fax: +44(0)20 7566 2334 20

Email: d.ezra@ucl.ac.uk 21

2

Disclosures; Drs. Keane, Rose and Ezra have received a proportion of their funding 22

from the Department of Health’s NIHR Biomedical Research Centre for 23

Ophthalmology at Moorfields Eye Hospital and UCL Institute of Ophthalmology. The 24

views expressed in the publication are those of the authors and not necessarily 25

those of the Department of Health. 26

Dr. Keane has received travel grants from the Allergan European Retina 27

Panel. 28

Dr. Timlin has no conflicts of interest. 29

30

Running Head: OCT imaging of the stenotic lacrimal punctum 31

32

33

34

Word Count: 2580 35

Figures: 4 36

Tables: 1 37

38

Abbreviations: 39

OCT – optical coherence tomography 40

EDI – enhanced depth imaging 41

ASM – anterior segment module 42

IR – infrared 43

HEYEX – Heidelberg Eye Explorer 44

NLDO – nasolacrimal duct obstruction45

3

ABSTRACT 46

Purpose: To determine the application of imaging the stenotic lacrimal punctum 47

with infrared photographs and optical coherence tomography (OCT), and to identify 48

characteristics of the lacrimal punctum in patients who benefit from punctoplasty. 49

Design: Case-Control Study 50

Subjects: 20 patients with epiphora who were listed for punctoplasty, and 20 healthy 51

controls. 52

Methods: Prospectively, 20 patients listed for punctoplasty were asked to rate their 53

epiphora, using the Munk score, before and after punctoplasty. They also underwent 54

pre-operative OCT and infrared imaging of the affected punctum. They were divided 55

into two groups, depending upon whether their epiphora improved or not, and were 56

compared to 20 healthy controls. 57

Main Outcome Measures: Measurements of puncta from infrared and OCT images 58

were taken along with Munk scores of patients’ epiphora. 59

Results: The infrared image measurements were significantly smaller in those 60

patients whose epiphora improved compared to those that did not, in both the area 61

of the punctal aperture and in the maximum punctal diameter. Additionally those 62

patients with improvement in epiphora had a significantly smaller preoperative 63

punctal diameter at 100μm depth on OCT, as compared to healthy controls; this was 64

not observed in patients whose epiphora failed to improve. There was no significant 65

difference in the punctum diameter between the three groups at the punctum surface 66

entrance or at 500μm depth. 67

Patients with epiphora had a higher tear meniscus within the punctum compared to 68

healthy controls. 69

4

Conclusions: Lacrimal punctum infrared and OCT imaging may be helpful in 70

predicting patients more likely to benefit symptomatically from punctoplasty, with 71

patients with smaller puncta have greater symptomatic improvement. However, it 72

also suggests that inner punctum diameter (not readily measurable by slit-lamp 73

examination), rather than the surface diameter, is correlated with outcome. 74

Additionally, OCT measurements of the tear meniscus height within the punctum 75

may be related to the degree of epiphora. 76

77

78

Keywords: Lacrimal punctum, stenosis, epiphora, optical coherence tomography, 79

non-invasive, infrared, ampulla, punctoplasty. 80

81

5

INTRODUCTION 82

The treatment of epiphora can be challenging for the clinician, especially 83

when nasolacrimal duct obstruction (NLDO) is excluded. If the lacrimal puncta are 84

thought to be site of critical flow resistance, punctoplasty is often undertaken to 85

enlarge the entrances to the drainage system. Punctoplasty can, however, be 86

ineffective and may rarely worsen epiphora due to impairment of the lacrimal pump 1, 87

or by scarring and contracture of the surgical opening. The uncertainty over outcome 88

after punctoplasty indicates a poor understanding of punctal dynamics, and in 89

identifying the clinical characteristics that may predict a beneficial outcome for the 90

procedure. 91

A large range of physiological punctal size and morphology allows adequate 92

physiological tear drainage 2,making it difficult to define whether a punctum is 93

“normal” or whether it represents an acquired stenosis that is amenable to 94

punctoplasty. An improvement in identifying those patients who are most likely to 95

benefit from punctoplasty is, therefore, beneficial. 96

Optical coherence tomography (OCT) is commonly used for examining the 97

retina and cornea. Only recently has it been used to image the lacrimal punctum and 98

to determine normal OCT characteristics in healthy individuals without epiphora 2-5. 99

OCT imaging is readily available in the ophthalmic clinic as a painless, non-invasive 100

and convenient imaging tool, and has been described as “in vivo clinical biopsy”6. 101

OCT can image up to 1.6mm depth 7 in skin, usefully imaging superficial tissue 102

layers. However, the semitransparent nature of the conjunctiva may allow deeper 103

penetration or greater detail in the peripunctal area. 104

OCT devices have rapidly evolved to deliver detailed tissue images. The initial 105

“time domain” OCTs developed into up to 100 times faster “spectral domain” OCTs 6. 106

6

Aggregation of multiple spectral domain OCT scans then provided “enhanced depth 107

imaging” (EDI) OCT8. Furthermore, customized anterior segment lenses, have led to 108

“swept source” OCT, enabling greater tissue penetration with the use of a longer 109

wavelength of light6. All of these advances in OCT mean that images of higher 110

resolution can be acquired in the lacrimal punctal area 2. 111

We report the application of spectral domain OCT with an anterior segment 112

module, and EDI scanning protocols, for in vivo assessment of the lacrimal puncta of 113

patients with punctal stenosis and listed for punctoplasty. The application of this 114

technique, and quantitative measurements of punctal size is provided for a cohort of 115

patients undergoing punctoplasty for epiphora. We compare the data from patients 116

whose symptoms improved after punctoplasty, and those whose symptoms did not, 117

with asymptomatic healthy volunteers 2. Recommendations will be made for punctal 118

OCT measurements in future research. 119

120

MATERIALS AND METHODS 121

Ethics 122

Regional Ethics Committee approval was obtained (LREC ref: 14/LO/1450; 123

153332 Westminster NRES Committee). The research adhered to the tenets of the 124

Declaration of Helsinki. 125

Subjects 126

Twenty patients listed for punctoplasty for treatment of epiphora were 127

recruited prospectively and written informed consent was obtained for the study. 128

Information regarding age, gender, and ethnicity, was obtained from all participants. 129

7

Patients were excluded if they did not have a lacrimal drainage system fully 130

patent to irrigation, or if they had undergone previous surgery to the eyelids or 131

lacrimal drainage system. 132

Pre-operative assessment of epiphora 133

Prior to undergoing punctoplasty, patients were asked to grade the degree of 134

epiphora on a typical day in each eye, using the Munk scoring system -- with a Munk 135

score of “0” for no epiphora, “1” for epiphora requiring dabbing with a tissue less than 136

twice a day, “2” for dabbing 2-4 times daily, “3” for dabbing 5-10 times daily, and “4” 137

for epiphora requiring dabbing more than ten times a day 9. 138

Image acquisition protocol. 139

OCT image-sets of both lower lacrimal puncta were obtained by a single 140

operator (H.T.) prior to punctoplasty, using a previously-described method 2. Briefly, 141

a single Spectralis OCT device with “Anterior Segment Module” (ASM) (Heidelberg 142

Engineering, Germany), which consists of an add-on lens and dedicated software, 143

was used. This system acquires 40,000 A-scans per second with a 7µm axial 144

resolution in tissue, and a transverse resolution of 14µm. All images for this study 145

were acquired using the scleral setting, a mode in which EDI-OCT can be performed. 146

Each cross-sectional image subtended an angle of 15º, covering an eyelid length of 147

approximately 8mm, and single scans were acquired with the automated real time 148

(ART). In order to align the punctum opening with the scan, the nasal lower eyelid 149

was gently rolled to evert the punctum (Figure1A). Simultaneous IR images were 150

taken by the Spectralis system and displayed alongside the OCT images (Figures 1A 151

and B). As the punctum was usually oval, the axis of the scan was rotated to align 152

with the largest diameter of the punctum aperture, usually parallel to the 153

mucocutaneous junction. 154

8

Image analysis 155

Images were analyzed using Heidelberg Eye Explorer (HEYEX) software 156

(version 1.6.8). All measurements were taken twice, by the same individual, with the 157

second performed without being able to see the location or size of the first 158

measurement, and then an average taken. All lines were drawn manually. 159

The infrared images provide 2 measurements: the area of the punctal 160

aperture (Figure 2A) and the maximum diameter of the opening (Figure 2B). A 161

change in greyscale was used to determine the darker punctum from the lighter lid 162

margin and a line was drawn at this change and the software generated the area 163

within this shape drawn. Following this, the maximum diameter of this shape was 164

hand generated and measured. 165

The OCT image aligned with the maximum punctal diameter on IR imaging 166

was used to derive 5 characteristics. The punctal diameter was measured at three 167

depths: at the external surface (Figure 3A), taken at the punctal opening by forming 168

a tangent connecting the highest points on the nasal and temporal punctal walls; the 169

second diameter was at 100μm below this surface measurement (Figure 3B); and 170

the third was at 500μm below the surface (Figure 3C and 3D). The other two OCT 171

characteristics were punctal depth (Figure 3E) and the depth of the intrapunctal tear 172

meniscus (Figure 3F), both measured from surface tangent. Four of the parameters 173

were taken as previously described 2; However, the punctal aperture at 100μm 174

depth was added because other papers have assessed punctal size from within the 175

opening, rather than at the surface 3-5; a depth of 100μm was chosen to reflect a 176

comparable intrapunctal measurement. 177

Patients were contacted by telephone 4-6 months following punctoplasty and 178

asked to rate their epiphora for each eye using the Munk score. Statistical analysis 179

9

with Student’s t-testing was performed using commercially available software 180

(Intercooled Stata for Windows, Version 9, Statacorp LP, USA).181

182

RESULTS 183

Patients underwent punctoplasty at a median age of 54 years (range 29-82), 184

and the median age of healthy volunteers was 37 years (range 27-64). Twelve (60%) 185

of the patients and 9 (45%) of the healthy controls were female. Eight (40%) patients 186

and 10 (50%) controls were Caucasian, 8 (40%) patients and 9 (45%) controls were 187

Asian, and 4 (20%) patients and 1 (5%) controls were Afro-Caribbean. 188

Half of the patients had a unilateral punctoplasty, and the other half had 189

bilateral disease. 190

Munk scores 191

Munk scores improved by a mean of 1.7 (range -1 to 4) after punctoplasty and 192

57% of treated puncta (17 puncta in 12 patients) had reduced epiphora, 37% (11 193

puncta of 7 patients) had no change in their epiphora, and 6.7% (2 puncta of 1 194

patient) had worse epiphora after surgery. 195

Infrared images 196

The two measurements from IR images were significantly smaller for patients 197

in whom epiphora improved, as compared to those who did not improve -- in both the 198

area of the punctal aperture (0.091mm2 (sd 0.064) vs 0.149 mm2 (sd 0.083); 199

p=0.04), and in maximum punctal diameter (367μm (sd 138) vs 547μm (sd 204); 200

p=0.02). These two IR measurements were also significantly smaller than controls 201

for patients in whom epiphora improved, with a punctal area of 0.091mm2 vs 202

0.137mm2 (sd 0.068) (p=0.02) and maximum punctal diameter of 367μm (sd 138) vs 203

10

500μm (sd 130) (p=0.006). There was no significant difference between patients in 204

whom epiphora did not improve, as compared with control subjects. 205

OCT Images: Measurement of punctal width. 206

The patients with stenotic puncta showed the same three distinct tissue layers 207

on OCT as reported for normal subjects.2 208

Patients in whom epiphora improved had a significantly smaller OCT punctal 209

width at 100μm depth as compared to control subjects (175μm (sd 124) vs 257μm 210

(sd 123); p=0.02) in contrast, the width at this depth was similar in patients without 211

improvement in epiphora (224μm (sd 157)) and control subjects. 212

Notably, there was no significant difference in punctal widths in the 3 groups 213

at the punctal surface (645μm (sd 150) for controls, 583μm (sd 169) for improvers 214

and 653μm (sd 164) for non improvers) or at the 500μm depth (50μm (sd 104) for 215

controls, 43μm (sd 55) for improvers and 35μm (sd 61) for non improvers). 216

OCT images: Punctal depth 217

There was no significant difference in the depth of punctum visible on OCT 218

between control subjects, patients in whom epiphora improved and those in whom 219

epiphora did not improve (543μm (sd 327), 557μm (sd 370), and 551μm (sd 344), 220

respectively). 221

In these near physiological conditions, the puncta appeared closed at a depth 222

of 500μm in 38% (15/40) control puncta, 47% (8/17) puncta where epiphora 223

improved, and 54% (7/13) puncta where epiphora failed to improve. 224

OCT images: Fluid level 225

A fluid level was seen on OCT within the punctum of 70% (28/40) healthy 226

puncta, 88% of (15/17) puncta where epiphora improved and 85% (11/13) puncta 227

where epiphora did not improve. The average depth of the fluid level was 192μm (sd 228

11

207) in controls, 78.6μm (sd 41.6) in puncta where epiphora improved and 73.9μm 229

(sd 44.6) in puncta where epiphora did not improve. The depth within the puncta 230

whose epiphora improved was significantly smaller than control subjects (p=0.04). 231

Although the puncta whose epiphora did not improve had a similar and slightly 232

smaller depth than improvers, they did not show statistical significance. 233

OCT images: Lacrimal ampullae 234

Although three control puncta showed ampullae on OCT (Figure 4), none of 235

the patients with epiphora showed an ampulla on OCT. 236

237

DISCUSSION 238

New findings. 239

This pilot study shows that the quick, non-invasive investigations of OCT and 240

IR imaging can be performed with punctum stenosis, and that patients with smaller 241

punctal area and maximum diameter on IR imaging are more likely to benefit from 242

punctoplasty. The results of the OCT measurements suggest, however, that the 243

important measure of punctal function is their reduced diameter just within the 244

punctum, rather than at its entrance -- with a significant difference (between the 245

punctoplasty success and control groups) in the punctal width at 100μm depth, but 246

not so at the punctal entrance or at 500μm depth. We propose that future OCT 247

studies should use measurements of punctal size at 100μm depth. 248

The intrapunctal meniscus occurred at a significantly greater depth in control 249

subjects as compared to patients whose epiphora improved, this possibly suggesting 250

a minor degree of distal resistance to flow in patients with epiphora – for example, 251

reduced canalicular pump function or nasolacrimal duct flow resistance, unable to be 252

detected with syringing. Interestingly, using OCT, measurements of the height of the 253

12

lower lid tear meniscus have shown significant reduction when patients become 254

epiphora-free after punctoplasty10. These OCT characteristics of the tear meniscus 255

on the lid margin and in the punctum could be further investigated for correlation with 256

the degree of epiphora or its aetiology; this, in turn, may help to more accurately 257

select patients for punctoplasty. 258

The average punctal width found at 100μm depth was 257μm in our healthy 259

controls, this being in accord with other reports of 215μm 4 and 247μm 3, and we 260

would suggest using this depth as a standardized location for measuring punctal 261

width – to thereby facilitate future comparative studies. 262

The IR maximal punctal diameter was smaller than the OCT punctum external 263

surface diameter. This is because the IR measurement was taken where a degree of 264

shade occurred, which is within the punctum and not at the very surface. The 265

measurements suggest that the IR maximal punctal diameter corresponded to a 266

depth within the punctum of between 0 and 100μm. 267

The horizontal canaliculus has not been visualized in any puncta with the 268

current OCT technology, suggesting that the full length of the vertical canaliculus has 269

not been imaged in a single scan. The largest depth seen was 1308μm in a control 270

patient where the punctum lumen was still open at this depth. It is challenging to 271

image the whole punctum depth, as this requires alignment of the OCT angle within 272

the vertical lumen throughout its length, along with stable lid eversion. 273

274

Epiphora score 275

The Munk score was used because it asks, with one simple question, the 276

patient to capture their epiphora on a typical day for them. There are other more 277

complex epiphora scoring systems available for example the ‘Lac-Q’ questionnaire 278

13

11, which asks 9 questions, some of which are not relevant to punctal stenosis such 279

as stickiness, and the ‘Impact of epiphora on vision-related quality of life’ 280

questionnaire 12 with 10 questions. However, it is not necessarily useful to ask 281

multiple questions when one will capture the severity, especially if multiple questions 282

ask about activities that the patient may only rarely undertake such as nighttime 283

driving. 284

285

Limitations and possible future applications for punctal imaging 286

This study is limited in the number of subjects examined, and by the large 287

range of punctal diameters in healthy subjects 2. The median age difference, 54 vs. 288

37 years, between the epiphora group and healthy control is also limitation of this 289

study. This was due to the collection of controls being done before collecting patients 290

with epiphora, in order to test whether OCT punctum imaging was feasible. Future 291

studies could try to age match controls in order to reduce the chance of any 292

difference being due to aging changes of the punctum. 293

The decision to undertake punctoplasty was also made by several physicians, 294

with no strict definition for “punctal stenosis” 13. The variation in punctal sizes in the 295

patients listed for punctoplasty merely underlines the need for a more objective 296

method of assessing which patients might benefit from surgery. 297

All measurement lines were hand drawn as currently there is no software 298

generated specifically for the punctum. However, this would be beneficial for 299

reducing human error and improving repeatability and comparison from patient to 300

patient or the same patient at different time points. 301

Due to the knowledge that tears flow through the upper punctum as well as 302

the lower 14, 15, imaging the upper punctum was attempted initially in control subjects. 303

14

However this was technically challenging due to difficulty in rotating the punctum 304

opening to face the OCT machine, in a direction half way between complete upper 305

lid eversion and the normal slightly posteriorly angled position of the upper lid 306

punctum. This is presumed to be due to the larger height of the upper lid tarsus. 307

Future development of OCT machines may perhaps enable greater angulation of the 308

scanning beam to include examination of the superior punctum. 309

By measuring aspects of punctal morphology which are not visible with slit 310

lamp biomicroscopy, IR and OCT imaging of the lacrimal puncta may help determine 311

which patients would benefit from punctoplasty; it might also help predict patients 312

with lacrimal pump failure or functional NLDO, in whom punctoplasty might be 313

ineffective. This latter group of patients may be identifiable from OCT parameters 314

such as punctal aperture, presence of an ampulla, or depth of the intrapunctal fluid 315

meniscus. 316

A future study might select patients with unilateral epiphora, to allow an intra-317

subject comparison of two systems and repeat the IR and OCT examinations post 318

operatively such as at 6 months. 319

320

321

322

15

LEGENDS 323

324

Figure 1. Images of each punctum were taken and displayed simultaneously using 325

Infrared (left panels) and Optical Coherence Tomography (OCT) (right panels). 326

These are the lower puncta of right eyes being everted by a cotton bud. The line 327

indicates the axis of the scan, which was rotated to align with the largest diameter of 328

the punctum aperture, usually parallel to the mucocutaneous junction. [A] shows a 329

narrow punctal entrance which is open to a deep depth, whereas [B] shows a wide 330

punctum which rapidly tapers into a narrow lumen. 331

332

16

Figure 2. The punctal dimensions derived from infrared images were (A) the area of 333

punctal aperture (within white circle) and (B) the maximum diameter of punctum 334

opening (white line). 335

336

Figure 3. The punctal dimensions derived from OCT images were; [A] the external 337

surface punctal aperture, taken at the opening of the punctum by forming a tangent 338

connecting the highest points on the nasal and temporal punctal walls. 339

[B] the punctal width at 100μm below the surface (white line), measured vertically 340

from the tangent at the lateral punctum wall and in parallel to the tangent (in grey). 341

[C] the punctal width at 500μm below the surface (white line), measured vertically 342

from the tangent (in grey) at the lateral punctal wall. In this patient’s right eye the 343

punctum appears “closed” at 500μm depth. 344

[D] the punctal width at 500μm below the surface (white line) in the left eye of a 345

different patient, measured vertically from the tangent (grey line) at the lateral 346

punctal wall, showing an open punctum that is able to be measured. 347

[E] the punctal depth (white line), measured from the deepest part of the punctum to 348

the tangent, along a vertical line, parallel to the lumen (grey line). 349

17

[F] the depth of the intrapunctal tear meniscus (white line), measured from the 350

deepest part of the fluid level to the tangent (grey line). 351

352

Figure 4. OCT images show [A] punctum where no ampulla is visible, [B] and [C] 353

where ampullae were visible and [D] where an ampulla was visible with a misaligned 354

scan through the conjunctiva. 355

18

356

Table 1. Demographics and punctal characteristics 357

358

REFERENCES 359

360

1. Chak M, Irvine F. Rectangular 3-snip punctoplasty outcomes: preservation of the 361

lacrimal pump in punctoplasty surgery. Ophthal Plast Reconstr Surg 2009;25(2):134-5. 362

2. Timlin HM, Keane PA, Day AC, et al. Characterizing the lacrimal punctal region 363

using anterior segment optical coherence tomography. Acta Ophthalmol 2016;94(2):154-9. 364

3. Wawrzynski JR, Smith J, Sharma A, Saleh GM. Optical coherence tomography 365

imaging of the proximal lacrimal system. Orbit 2014;33(6):428-32. 366

4. Kamal S, Ali MJ, Ali MH, Naik MN. Fourier Domain Optical Coherence 367

Tomography With 3D and En Face Imaging of the Punctum and Vertical Canaliculus: A Step 368

Toward Establishing a Normative Database. Ophthal Plast Reconstr Surg 2016;32(3):170-3. 369

19

5. Allam RS, Ahmed RA. Evaluation of the Lower Punctum Parameters and 370

Morphology Using Spectral Domain Anterior Segment Optical Coherence Tomography. J 371

Ophthalmol 2015;2015:591845. 372

6. Drexler W, Fujimoto JG. State-of-the-art retinal optical coherence tomography. Prog 373

Retin Eye Res 2008;27(1):45-88. 374

7. Vokes DE, Jackson R, Guo S, et al. Optical coherence tomography-enhanced 375

microlaryngoscopy: preliminary report of a noncontact optical coherence tomography system 376

integrated with a surgical microscope. Ann Otol Rhinol Laryngol 2008;117(7):538-47. 377

8. Spaide RF, Koizumi H, Pozzoni MC, Pozonni MC. Enhanced depth imaging spectral-378

domain optical coherence tomography. Am J Ophthalmol 2008;146(4):496-500. 379

9. Munk PL, Lin DT, Morris DC. Epiphora: treatment by means of dacryocystoplasty 380

with balloon dilation of the nasolacrimal drainage apparatus. Radiology 1990;177(3):687-90. 381

10. Kim SE, Lee SJ, Lee SY, Yoon JS. Outcomes of 4-snip punctoplasty for severe 382

punctal stenosis: measurement of tear meniscus height by optical coherence tomography. Am 383

J Ophthalmol 2012;153(4):769-73, 73.e1-2. 384

11. Mistry N, Rockley TJ, Reynolds T, Hopkins C. Development and validation of a 385

symptom questionnaire for recording outcomes in adult lacrimal surgery. Rhinology 386

2011;49(5):538-45. 387

12. Shin JH, Kim YD, Woo KI, (KSOPRS) KSoOPaRS. Impact of epiphora on vision-388

related quality of life. BMC Ophthalmol 2015;15:6. 389

13. Kashkouli MB, Nilforushan N, Nojomi N, Rezaee R. External lacrimal punctum 390

grading: reliability and interobserver variation. Eur J Ophthalmol 2008;18(4):507-11. 391

14. Daubert J, Nik N, Chandeyssoun PA, el-Choufi L. Tear flow analysis through the 392

upper and lower systems. Ophthal Plast Reconstr Surg 1990;6(3):193-6. 393

20

15. White WL, Glover AT, Buckner AB, Hartshorne MF. Relative canalicular tear flow 394

as assessed by dacryoscintigraphy. Ophthalmology 1989;96(2):167-9. 395

396